1. Field of the Technology
The present disclosure relates generally to handover methods and apparatus between wireless local area networks (WLANs) and wireless wide area networks (WWANs) for mobile communication devices.
2. Description of the Related Art
The present disclosure relates generally to handover methods and apparatus between heterogeneous wireless networks, such as WLANs (e.g. IEEE 802.11 based networks) and WWANs (e.g. cellular telecommunication networks), for mobile communication devices. The specific problem addressed relates to the support of real-time voice calls (or other media communications) when “multi-mode” devices are utilized in enterprises network environments.
In such environments, each mobile device has a WLAN radio interface and a cellular radio interface. When a voice call is active via the WLAN radio interface and the mobile device roams out of WLAN coverage, the voice call is maintained by reconnecting it through the cellular radio interface of the mobile device. The transition between these two radio interfaces and networks is referred to as a vertical handover (VHO). The switch between interfaces must by done subject to strict latency constraints, so that the voice connection quality is not adversely affected.
In order to properly provide a VHO, the connection is normally split into two call “legs” which are anchored either in the cellular network or in the enterprise. The “anchor” is the point where the two call legs come together. Then VHO occurs, one of these legs is replaced by a new call leg that is established through the wireless network (WLAN or cellular) to which the mobile device is handing over. Enterprise anchoring (EA) is complex from a user's point of view since the handover must be anchored and managed by equipment inside the enterprise, such as a Public Switched Telephone Network (PSTN) gateway or an IP Public Branch Exchange (PBX). Cellular network anchoring (CNA) pushers this complexity into the cellular network, which is more desirable from that point of view. CNA is often capable of much faster handovers since both WLAN and cellular call legs terminate inside the cellular operator's core network. The CNA mode is typical of currently proposed carrier-based dual-mode device solutions such as IP Multimedia Subsystem (IMS) and Unlicensed Mobile Access (UMA). Enterprise anchoring normally incurs longer VHO execution delays than CNA because the new cellular call leg setup must propagate through the cellular core network, the PSTN, and the enterprise network.
The following documents are related to and may be referenced in the remaining discussion: [1] ETSI. “Requirements And Architectures For Interworking Between HIPERLAN/3 And 3rd Generation Cellular Systems”. Technical Report TR 101 957, ETSI, August 2001; [2] http://www.umatechnology.org, 2005; [3] R. Katz M. Stemm. “Vertical handoffs in wireless overlay networks”. Mobile Networks and Applications 3, pp. 335-350, 1998; [4] H. Choi et al. “A Seamless Handoff Scheme For UMTS-WLAN Interworking”. In GLOBECOM '04, pp. 1559-1564, vol 3, November-December 2004; [5] K. El Malki et al. “Low Latency Handoff In Mobile IPv4”, draft-ietfmobileip-lowlatency-handoffs-v4-01, IETF, May 2006; [6] R. Chakravorty et al. “Performance Issues With Vertical Handovers—Experiences From GPRS Cellular And WLAN Hot-Spots Integration”. In PERCOM '04, pp. 155-164, March 2004; [7] C. E. Perkins et al. “Route Optimization In Mobile IP”. Mobile IP working group, Internet draft-Work in process, November 1997; [8] C. E. Perkins et al. “Optimized Smooth Handoffs In Mobile IP”. In IEEE International Symposium on Computers and Communications, pages 340-346, July 1999; [9] T. Adachi and M. Nakagawa. Capacity Analysis For A Hybrid Indoor Mobile Communication System Using Cellular And Ad-Hoc Modes. In The 11th IEEE International Symposium on Personal Indoor and Mobile Radio Communications (PIMRC '2000), vol. 2, pp. 767-771, 2000; [10] R.-S. Chang, W.-Y. Chen, and Y.-F. Wen, “Hybrid Wireless Network Protocols”, IEEE Transactions on Vehicular Technology, 52(4):1099-1109, July 2003; [11] X. Wu, S-H. G. Chan, and B. Mukherjee. MADF: A Novel Approach To Add An Ad-Hoc Overlay On A Fixed Cellular Infrastructure. In IEEE Wireless Communications and Networking Conference (WCNC '2000), vol. 2, pp. 549-554, 2000; [12] C. Qiao and H. Wu, iCAR: An Intelligent Cellular And Ad-Hoc Relay System. In Ninth International Conference on Computer Communications and Networks, pages 154-161, 2000; [13] Y.-D. Lin and Y.-C. Hsu, Multihop Cellular: A New Architecture For Wireless Communications. In IEEE INFOCOM 2000, vol. 3, pp. 1273-1282, 2000; [14] B. S. Manoj R. Ananthapadmanabha and C. S. R. Murthy. Multi-Hop Cellular Networks: The Architecture And Routing Protocols. In 12th IEEE International Symposium on Personal, Indoor and Mobile Radio Communications, vol. 2, pp. G78-G82, 2001; [15] T. Rouse, I. Band, and S. McLaughlin. Capacity And Power Investigation Of Opportunity Driver Multiple Access (ODMA) Networks In TDD-CDMA Based Systems. In IEEE International Conference on Communications, 2002; [16] G. N. Aggelou and R. Tafazolli. On The Relaying Capability Of Next-generation GSM Cellular Networks. In IEEE Personal Communications, pp. 40-47, February 2001; [17] J. H. Yap, X. Yang, S. Gbaheri-Niri, and R. Tafazolli. Position Assisted Relaying And Handover In Hybrid Ad Hoc WCDMA Cellular System. In 13th IEEE International Symposium on Personal, Indoor, and Mobile Radio Communications. (PIMRC '2002). Lisbon, Portugal, pp. 2194-2198, September 2002; [18] E. Balafoutis A. Panagakis and I. Stavakskis, “Study Of The Capacity Of Multihop Cellular Networks”. Lecture Notes in Computer Science, Springer-Verlag, Ad Hoc Assisted Handoff In IEEE 802.11 Infrastructure Networks. International Journal on Computer and Communication Networks, to appear, 2005; [20] J. Rosenberg et al. “SIP: Session Initiation Protocol”. RFC 3261, 2002; [21] www.ovislink.ca/wls/gp1000.htm; [22] www.sourceo2.com/o2 developers/o2 technologies/gprs/technical overview; [23] B. Goodman. “Internet Telephony And Modem Delay”. IEEE Network, 13(3); 8-16, May-June 1999; [24] H. Velayos et al. “Techniques To Reduce The IEEE 802.11b Handoff Time”. In IEEE International Conference on Communications vol. 7, pages 3844-3848, June 2004; and [25] www.techonline.com/community/edresource/14365.
With respect to such concerns, The European Telecommunications Standards Institute (ETSI) has specified two approaches for WLAN/WWAN interworking, namely, loose coupling approach and tight coupling approach (see reference [1] above). In the tight coupling approach, the WLAN is integrated into the service provider's cellular core network. An interworking gateway provides adaptation between the two systems. Tight coupling benefits from good handover delay and packet loss performance, as it uses the native cellular mobility management protocols. This approach is currently being standardized under the UMA or 3GPP-GAN activities (see reference [2] above). In contrast, a loosely-coupled approach connects the WLAN to the cellular network through an external IP network. This approach is potentially more scalable and less proprietary, but real-time handover may be more difficult to achieve. The loosely-coupled approach is the preferable choice for techniques of the present disclosure.
A VHO between difference radio interfaces involves a variety of time-consuming procedures such as handover triggering, base station selection, authentication, service negotiation, and IP address acquisition. As the procedures are very time consuming, they can significantly disrupt real-time (e.g. voice) communication. This is particularly true when handing off in the WLAN-to-cellular direction, as WLAN signal degradation can be very abrupt and unexpected. In many common situations, such as when exiting a building during an active connection, WLAN coverage loss may occur with very little warning at all. Even if the cellular interface of the mobile device is active at the time of transition, considerable packet loss will typically occur before the connection is recovered.
Some of the first work on VHO was done as documented in reference [3] above, which used a combination of analytic models and testbed experiments, proposing to make use of a multicast address in the mobile device which received advertisements from potential access points in an overlay. Handover initiation relies on the detection of periodic beacons from the different networks. It was shown that handover latencies can easily be as high as three (3) seconds. Fast beaconing and packet/header double-casting may reduce this delay to 800 msec (see reference [3] above). In reference [4] above, a smooth VHO scheme using pre-authentication and pre-registration was proposed for WLANs tightly coupled to a UMTS network. Pre-registration is a mobile IP-based fast handover scheme that triggers MIP handover before link layer handover, thereby limiting packet loss and handover delay to that caused by link layer handover (see reference [5] above). Moreover, by having the old AP buffer data packets during handover, data packet loss is further reduced. This forwarding mechanism is reasonable as the APs involved are typically separated by a small number of forwarding hops. This is usually not the case in a loosely-coupled WLAN/cellular architecture. Another study as documented in reference [6] above has experimented with a loosely-coupled MIPv6-based GPRS-WLAN testbed and has investigated the impact of VHO on TCP connections. Their experiments indicate a 3.8 sec VHO delay, and that by using fast router advertisements (RA), RA Caching, and binding update simulcasting, it may be reduced to about 1.36 sec.
In reference [7] above, a route optimization extension has been proposed for Mobile IP using the binding update message. A binding update can also be used to provide soft handover. Before completing the registration process and before the flow of data packets through the new foreign agent (NFA) starts, the mobile device requests that the NFA send a binding update to the old FA. The OFA then realized the current IP address for the mobile device and forwards the data packets still arriving on the old path to the current location of the mobile device hence reducing packet loss. In another improvement to Mobile IP, documented in reference [8] above, a buffering an forwarding scheme is proposed at the foreign agents (FA) to reduce data loss during a handover. The FA buffers any data packet it is forwarding to the mobile device. When a handover occurs, the new FA requests the old FA to forward buffered data packets. The new FA in turn forwards these data packets to the mobile device. The idea of FA forwarding has also been presented in the context of a post-registration handover scheme which is an extension to Mobile IP (see reference [5] above). To take advantage of these proposals the foreign agents and the MIP protocol needs to be changed. Also, the forwarding is done through the wired network which is likely to be very lengthy in WLAN-to-cellular handover cases.
There has been a lot of recent activity that considers the inclusion of ad hoc relaying into wireless infrastructure networks. A variety of systems have been considered, which often differ on the basis of whether mobile devices have multiple air interfaces, whether ad hoc infrastructure is present, and whether WLAN and/or cellular is being considered. The system described in reference [9] above uses ad hoc networking to enable communications whenever nodes are within range without using the cellular infrastructure. This is also the objective in reference [10] above, but to maintain simplicity a maximum of two ad hoc hops may be used between the end stations. In Mobile Assisted Data Forwarding (MADF) described in reference [11] above, special forwarding channels are allocated from resources used by the existing cellular network. These channels are then used for relaying traffic between cells. A mobile device which is about to handover finds a mobile device within range and link quality with both itself and the AP and requests relaying. This extends the coverage of the current cell and provides time for the mobile device to complete its handover. The approach in Integrated Cellular Ad Hoc Relaying System (ICAR) described in reference [12] above is similar to this approach, but uses special pre-installed ad hoc relay stations (RSs) to move traffic between cells. The multi-hop cellular system incorporates ad hoc routing into the cellular network using the same air interface as that used by the cellular base stations (see references [13] and [14] above). This concept is similar to the opportunity driven multiple access (ODMA) system proposed in reference [15] above and the system described in reference [16] above.
In reference [17] above, a technique referred to as position-assisted relaying was proposed for WCDMA cellular networks with dual-mode stations. In this scheme, a nearby station may relay transmissions for another when that station's cellular link becomes unusable. Geo-location techniques such as GPS or OTDoA are used by the base station to select a candidate RS (see reference [17] above). In reference [18] above, ad hoc relaying was studied from a capacity viewpoint in an IEEE 802.11 network. This study showed that inband relaying can significantly degrade access point capacity due to interference effects. A multi-frequency approach is better from this point of view. In reference [19] above, ad hoc relaying was used in infrastructure-based IEEE 802.11 networks that are supporting real-time voice connections. In such a system, active voice calls must be handed off between access points when the mobile device passes from one wireless coverage area to another. In this case, relaying is used to extend WLAN coverage when a mobile device moves outside the range of an AP.
Again, a VHO between different radio interfaces involves a variety of time-consuming procedures which may significantly disrupt real-time (e.g. voice) communication. This is particularly true when handing off from the WLAN-to-WWAN direction, as WLAN signal degradation can be very abrupt and unexpected. In many situations, e.g. when exiting a building during an active connection, WLAN coverage loss may occur with very little warning at all. Even if the cellular interface of the mobile device is active at the time of transition, considerable packet loss will typically occur before the connection is recovered.
Accordingly, there is a need for improved WLAN-to-WWAN handover methods and apparatus that overcome the deficiencies of the prior art.